The complement receptor C5aR2 promotes protein kinase R expression and contributes to NLRP3 inflammasome activation and HMGB1 release from macrophages

The NLR family pyrin domain-containing 3 (NLRP3) inflammasome is a multimeric protein complex that mediates maturation of the cytokines IL-1β and IL-18 as well as release of the proinflammatory protein high-mobility group box 1 (HMGB1) and contributes to several inflammatory diseases, including sepsis, gout, and type 2 diabetes. In this context, the well-studied active complement fragment C5a and its receptor C5aR1 or C5aR2 orchestrate the inflammatory responses in many diseases. Although a C5a-C5aR interaction in NLRP3-associated diseases has been suggested, little is known about the details of C5a–C5aR cross-talk with the NLRP3 inflammasome in macrophages. In this study, using mice and murine macrophages and cytokines, immunoblotting, siRNA, and quantitative real-time PCR assays, we demonstrate that C5aR2 deficiency restricts activation of the NLRP3 inflammasome and release of HMGB1 both in vitro and in vivo. Mechanistically, we found that C5aR2 promotes NLRP3 activation by amplifying dsRNA-dependent PKR expression, which is an important NLRP3-activating factor. We also observed that elevation of PKR expression because of the C5a–C5aR2 interaction depends on the mitogen-activated protein kinase/extracellular signal-regulated kinase kinase pathway and type I IFN signaling. In conclusion, these findings reveal that C5aR2 contributes to NLRP3 inflammasome activation and HMGB1 release from macrophages.


NLRP3 inflammasome activation and HMGB1 release from macrophages.
Inflammasomes are large multimeric protein complexes composed of a cytosolic sensor (e.g. NLRP3), an adaptor protein (apoptosis-associated speck-like protein containing a CARD (ASC)), and an effector, procaspase-1, which play a vital role in host defense against pathogens and inflammation (1,2). Among these inflammasomes, the NLRP3 inflammasome has received much attention. A wide variety of pathogen-associated molecular patterns, such as nigericin (Nig), 4 and damage-associated molecular patterns, including ATP and monosodium urate (MSU) crystals, can activate the NLRP3 inflammasome and lead to maturation of IL-1␤ and IL-18 (3). In addition, accumulated evidence shows that the NLRP3 inflammasome also mediates the release of high-mobility group box 1 (HMGB1), a late mediator of lethal sepsis, and contributes to the pathogenesis of septic shock (4 -6). It has been well-documented that the NLRP3 inflammasome is involved in many inflammatory diseases and disorders, including sepsis, gout, type 2 diabetes, colitis, atherosclerosis, and arthritis (7)(8)(9)(10). Consequently, proper regulation of the NLRP3 inflammasome is critical for the maintenance of immune homeostasis.
The active complement peptide C5a has been recognized as a powerful proinflammatory mediator, interacting with its receptors (C5aR1 and C5aR2) after onset of sepsis (11,12). There are numerous studies suggesting that interaction of C5a with its receptors is linked to many inflammatory diseases, such as sep-sis, septic cardiomyopathy, and gout (13)(14)(15)(16). All of these studies indicate that blocking C5a or absence of the C5a receptor could limit inflammatory responses, decrease IL-1␤, or suppress neutrophil recruitment and provide a protective role in such diseases (13)(14)(15)(16)(17). In addition, C5a-C5a receptor interaction in NLRP3-associated diseases has been suggested. However, the cell types used in these studies were cardiomyocytes (14,15) rather than immune cells, and the underlying mechanism of C5a in NLRP3 inflammasome activation needs to be explored further. Recent studies have indicated that C5a-C5aR1 regulates lipopolysaccharide (LPS)-induced NLRP3 inflammasome activation in monocytes (18); therefore, we asked whether C5a-C5aR2 has similar effects on the modulation of NLRP3 inflammasome activation in macrophages and the associated disease models.
In this study, we demonstrated that C5aR2 deficiency dampens activation of the NLRP3 inflammasome and the release of HMGB1 in vitro and in vivo. Mechanistically, C5aR2 promotes NLRP3 activation by amplifying PKR expression, which is a well-studied NLRP3 activating factor. Moreover, C5a-C5aR2promoting PKR expression depends on the MEK/ERK pathway and type I IFN signaling. Thus, our study reveals that C5aR2 contributes to NLRP3 inflammasome activation and HMGB1 release in macrophages.

C5aR2 deficiency inhibits NLRP3 inflammasome activation and the release of HMGB1
To determine the role of C5aR2 in NLRP3 inflammasome activation, LPS-primed mouse peritoneal macrophages from C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mice were treated with NLRP3 agonists: ATP, nigericin, and MSU. We found that C5aR2 deficiency significantly reduced IL-1␤ production but not TNF␣ production (Fig. 1A). In addition, cleavage of caspase-1 and IL-1␤ was dampened by C5aR2 deficiency, whereas expression of NLRP3 was not affected by C5aR2 (Fig. 1B), and NLRP3 is required for IL-1␤ maturation in this process (Fig. S1, A-D). Previous research has suggested that C5aR2 is required for the release of HMGB1 (13). We therefore assessed the abundance and found that NLRP3-induced HMGB1 was also decreased in the absence of C5aR2 (Fig. 1C). Early studies indicated that NLRP3 is critical for IL-1␤ processing via noncanonical inflammasome activation (8), so we examined the role of C5aR2 in noncanonical inflammasome activation and obtained similar results as for canonical inflammasome activation (Fig. 1, D-F).
Previous studies (19 -21) have demonstrated the importance of C5aR1 in inflammatory responses, so we compared the contribution of C5aR1 and C5aR2, individually and collectively, to the activation of NLRP3 inflammasomes. We applied siRNAs to silence C5aR1 in C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mouse macrophages, followed by stimulation with NLRP3 agonists. The knockdown efficiency of C5aR1 is more than 70% (Fig. S2, A  and B). The results showed that knockdown of C5aR1 in WT cells barely affects NLRP3 inflammasome activation. Further, interference of C5aR1 in C5aR2 Ϫ/Ϫ macrophages (a mimic of a double defect of C5aR1 and C5aR2) had a similar effect on the production of IL-1␤ compared with C5aR2 Ϫ/Ϫ cells (Fig. S2C). This is in accordance with a previous study in which C5aR1 deficiency had no effect on NLRP3 inflammasome activation (18) and also suggests a possible dominant role of C5aR2 rather than C5aR1 in activation of NLRP3. To determine whether C5aR2 specifically regulates the NLRP3 inflammasome, we stimulated LPS-primed macrophages from C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mice with poly(dA:dT) (an AIM2 inflammasome activator), flagellin (an NLRC4 inflammasome activator), and nigericin. The abundance of IL-1␤ induced by AIM2 and NLRC4 agonists was slightly decreased by C5aR2 deficiency (Fig. S3A), whereas there was no difference in production of TNF␣ (Fig. S3B). Additionally, we checked C5aR2's role in TLR signaling (TLR4 stimulated by LPS, TLR2 stimulated by Pam3CSK4, and TLR3 stimulated by poly(I:C)) and found that TNF␣ or IL-6 was barely influenced by C5aR2 deficiency (Fig.  S3, C and D). Taken together, these data suggested that C5aR2 mainly contributes to activation of the NLRP3 inflammasome and the release of HMGB1.

C5aR2 deficiency inhibits NLRP3 inflammasome activation in vivo
To further investigate the function of C5aR2 in NLRP3 inflammasome activation in vivo, C5aR2 Ϫ/Ϫ and C5aR2 ϩ/ϩ mice were injected with MSU to induce NLRP3 inflammasomedependent peritoneal inflammation. The IL-1␤ level and peritoneal exudate cell numbers in the peritoneal lavage fluids were measured. Compared with C5aR2 ϩ/ϩ mice, C5aR2 Ϫ/Ϫ mice showed an obvious reduction in peritoneal inflammation (Fig.  2, A and B). Additionally, we used a LPS-induced sepsis model to examine the effect of C5aR2. In response to LPS challenge, the serum levels of IL-1␤ and HMGB1, but not TNF␣, were greatly reduced in C5aR2 Ϫ/Ϫ mice compared with C5aR2 ϩ/ϩ mice (Fig. 2C). C5aR2 Ϫ/Ϫ mice also exhibited higher resistance to LPS-induced lethality than C5aR2 ϩ/ϩ mice (Fig. 2D), which was consistent with an earlier report that C5aR2 Ϫ/Ϫ mice have improved survival compared with C5aR2 ϩ/ϩ mice after cecal ligation and puncture (CLP)-induced sepsis (13). Collectively, these data suggested that C5aR2 is needed for activation of the NLRP3 inflammasome in vitro and in vivo.

C5aR2 deficiency reduces the expression of PKR in macrophages
We and others have suggested that PKR is required for NLRP3 inflammasome activation and HMGB1 release (22)(23)(24)(25). Thus, we wanted to find out whether PKR is involved in C5aR2-mediated NLRP3 inflammasome activation and HMGB1 release. Accordingly, we examined the expression and the phosphorylation level of PKR in NLRP3 agonisttreated macrophages from C5aR2 Ϫ/Ϫ and C5aR2 ϩ/ϩ mice. Surprisingly, expression and the level of PKR phosphorylation were diminished in the absence of C5aR2 (Fig. 3, A and  B). This gave us a clue to further explore the relationship between C5aR2 and PKR. Then we examined the expression of PKR in macrophages treated with LPS, the C5aR2 natural ligand C5a, or both. We found that C5a alone (1 g/ml) could promote the expression of PKR as much as LPS ( Fig.  3C and D). Although combining with LPS, C5a promoted the expression of PKR and increased the phosphorylation level C5aR2 promotes NLRP3 activation of PKR in a dose-dependent manner ( Fig. 3C and D). Moreover, LPS or C5a or the both slightly promoted PKR expression in macrophages in the absence of C5aR2. (Fig. 3C and  D). Thus, C5aR2 deficiency reduces the expression of PKR as well as phosphorylation of PKR in macrophages upon LPS or C5a stimulation.

PKR rescues NLRP3 inflammasome activation in the absence of C5aR2
We tested whether PKR could rescue NLRP3 inflammasome activation in the absence of C5aR2. We electroporated PKR or an empty vector to macrophages from C5aR2 ϩ/ϩ and A-C, after priming with LPS (100 ng/ml) for 3 h, peritoneal macrophages collected from C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mice were treated with NLRP3 agonists: ATP (5 mM for 1 h), nigericin (10 M for 1 h), and MSU (200 g/ml for 6 h). Supernatants (Sup) were analyzed by ELISA for IL-1␤ and TNF␣ (A) and HMGB1 release (C). B, after stimulation with canonical NLRP3 agonists as in A, release of HMGB1 in supernatants, cleavage of caspase-1 and IL-1␤, and expression of pro-caspase-1, pro-IL-1␤, and NLRP3 in cell extracts were assayed by Western blotting. Ctrl, control. D-E, peritoneal macrophages collected from C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mice were treated with noncanonical NLRP3 agonists: LPS transfection (2 g/ml of LPS transfected with 0.25% of FuGENE HD) and CTBϩLPS (2 g/ml of LPS combined with 5 g/ml of CTB). Supernatants were analyzed by ELISA for HMGB1 (D) and IL-1␤ and TNF␣ release (E). F, after stimulation with noncanonical NLRP3 agonists as in D, release of HMGB1 in supernatants, cleavage of caspase-1 and IL-1␤, and expression of pro-caspase-1, pro-IL-1␤, and NLRP3 in cell extracts were measured by Western blotting. The results represent the mean Ϯ S.D. of three independent experiments performed in triplicate (A and C-E). Two-tailed Student's t test was used (A and C-E). *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; ns, not significant. For Western blot analysis, data are representative of three independent experiments (B and F).
C5aR2 Ϫ/Ϫ mice and then monitored the expression of IL-1␤ and HMGB1 after stimulation with NLRP3 agonists. The transfection efficiency was quantified by real-time qPCR (Q-PCR) (Fig. 4A), and the ELISA results showed that PKR transfection could partially rescue IL-1␤ and HMGB1 release in the absence of C5aR2 (Fig. 4, B and C). These data demonstrate that PKR is pivotal for C5aR2-mediated NLRP3 inflammasome activation and HMGB1 release.  A and B, after priming with LPS (100 ng/ml) for 3 h, C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mouse peritoneal macrophages were treated with canonical (LPSϩATP, LPSϩNig, or LPSϩMSU) or noncanonical (LPS transfection or CTBϩLPS) NLRP3 agonists. Total RNA was extracted and subjected to Q-PCR (A), and total protein was extracted and subjected to Western blot analysis of PKR expression and the phosphorylation level of PKR (B), respectively. Ctrl, control. C and D, peritoneal macrophages from C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mice were treated with LPS (100 ng/ml) and/or increasing concentrations of recombinant C5a protein (0.1, 0.5, and 1 g/ml) for 4 h, and then total RNA (C) and protein (D) were extracted and subjected to Q-PCR and Western blot analysis of PKR expression and the phosphorylation level of PKR, respectively. The results represent the mean Ϯ S.D. of three independent experiments performed in triplicate (A and C). Two-tailed Student's t test was used (A and C). For Western blot analysis, data are representative of three independent experiments (B and D). *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001.

C5aR2 promotes NLRP3 activation C5a induces proinflammatory mediators via PKR in vivo
To further verify the significance of PKR in C5a-mediated inflammation in vivo, we injected LPS plus recombinant C5a to PKR ϩ/ϩ and PKR Ϫ/Ϫ mice and collected whole blood. In PKR ϩ/ϩ mice, the levels of serum IL-1␤ and HMGB1 from the LPS plus C5a-treated group were significantly higher than in the LPS alone-treated group, whereas in PKR Ϫ/Ϫ mice, there was no statistically significant difference between the two groups (Fig. 5, A and B), which further demonstrated that PKR regulates the production of C5a-induced proinflammatory mediators in vivo. These results indicate that the C5a-C5aR2 axis promotes NLRP3 activation and HMGB1 release by amplifying the expression of PKR.

C5a promotes the expression of PKR through MEK/ERK signaling and the type I IFN pathway
To determine the potential signal pathway underlying the C5aR2-mediated promotion of PKR expression, we used only C5a to stimulate the macrophages, as it is the natural ligand for C5aR2 and has a similar promoting effect with LPS. We examined MEK/ERK, mitogen-activated protein kinase 14 (p38), and PI3K regulatory subunit 1/Akt, which have been reported to mediate C5a's function in innate immune cells (18,26,27). After treatment with a MEK1/2 inhibitor (U0126), a p38 inhibitor (SB203580), or a PI3K inhibitor (wortmannin), macrophages were challenged with C5a. Our results indicated that the MEK1/2 inhibitor greatly reduced the expression of PKR induced by C5a, whereas the other two had no significant effect (Fig. 6A). The specificity of the inhibitors was detected according to previous research (28 -33) (Fig. S4, A-C). In addition, IL-1␤ secretion was decreased when U0126 or SB203580 was added to macrophages under LPSϩNig stimulation (Fig. 6B). We also noticed that SB203580 could impair TNF␣ expression, whereas U0126 had no effect (Fig. 6B), which is because SB203580 could suppress NF-B activation (33)(34)(35). To further validate the above signaling pathways, we synthesized siRNAs targeting MEK1, p38, and PI3K before C5a treatment. Only siMEK1 could reduce the expression of PKR following C5a stimulation (Fig. S5, A-D). Accordingly, MEK/ERK signaling is likely involved in C5aR2-mediated NLRP3 inflammasome activation.
Type I IFN signaling has been reported to regulate PKR expression (36), so we used type I IFN receptor knockout (IFN␣␤R Ϫ/Ϫ ) mice to further the investigation, in which type I IFN signaling is blocked. We stimulated macrophages from IFN␣␤R ϩ/ϩ and IFN␣␤R Ϫ/Ϫ mice with C5a protein. PKR expression was markedly decreased in IFN␣␤R Ϫ/Ϫ mice after C5a stimulation compared with IFN␣␤R ϩ/ϩ mice (Fig. 6C), suggesting that IFN signaling is involved in C5a-mediated promotion of PKR expression. Moreover, C5a treatment could not increase the production of IL-1␤ and TNF␣ in IFN␣␤R Ϫ/Ϫ macrophages upon NLRP3 agonist exposure (Fig. 6D). Overall, these data suggest that C5a-C5aR2 promotes PKR expression via MEK/ERK signaling and the type I IFN pathway.   A, WT mouse peritoneal macrophages were pretreated with U0126 (a pharmacologic MEK1/2 inhibitor), SB203580 (a selective p38 inhibitor), and wortmannin (a selective PI3K inhibitor) 1 h prior to C5a (1 g/ml) stimulation (3 h), and the PKR mRNA level was assayed by Q-PCR. B, after pretreatment with inhibitors as in A, WT mouse macrophages were primed with LPS (100 ng/ml) with or without C5a (1 g/ml) for 3 h, followed by stimulation with nigericin (10 M) for 1 h, and then supernatants were analyzed by ELISA for IL-1␤ and TNF␣. C, peritoneal macrophages from IFN␣␤R ϩ/ϩ and IFN␣␤R Ϫ/Ϫ mice were treated with C5a (1 g/ml) for 3 h, and PKR mRNA was detected via Q-PCR. D, peritoneal macrophages from IFN␣␤R ϩ/ϩ and IFN␣␤R Ϫ/Ϫ mice were treated with LPS (100 ng/ml) with or without C5a (1 g/ml) for 3 h, followed by stimulation with nigericin (10 M) for 1 h, and then supernatants were analyzed by ELISA for IL-1␤ and TNF␣. The results represent the mean Ϯ S.D. of three independent experiments performed in triplicate. Two-tailed Student's t test was used. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; ns, not significant.

Discussion
The anaphylatoxin C5a is generated upon complement activation and contributes to the development of many NLRP3associated inflammatory disorders, including sepsis (37), septic cardiomyopathy (15), and gout (38). However, little is known about C5a-C5a receptor cross-talk with the NLRP3 inflammasome in immune cells such as macrophages. In this study, we showed that, in the absence of C5aR2, macrophages produce less IL-1␤ and HMGB1 upon NLRP3 agonist treatment. Moreover, in the MSU-induced peritonitis model and LPS-induced sepsis model, C5aR2 Ϫ/Ϫ mice exhibited obviously reduced inflammation. To further uncover the mechanism, we surprisingly found that C5aR2 promotes the expression of PKR, an NLRP3 inflammasome activator, which was identified by our previous research (22). Furthermore, PKR transfection rescued inhibition of IL-1␤ and HMGB1 secretions in the presence of NLRP3 agonists and C5aR2 deficiency. In addition, PKR-deficient mice show attenuated production of IL-1␤ and HMGB1 after LPS plus C5a injection intraperitoneally compared with WT mice. Last, we found that C5a elevated PKR expression via MEK/ERK signaling and type I IFN signaling. Thus, our study revealed a previously unknown role of C5aR2 in NLRP3 inflammasome activation in macrophages (Fig. 7).
C5a-C5a receptor interaction has attracted much attention in inflammatory disorders as studies reveal their proinflammatory roles (38,39). An initial study found that, in a CLP-induced sepsis model, C5aR1-or C5aR2-deficient mice had improved survival compared with WT mice and that C5aR2 is required for HMGB1 release (13). Based on this study, our work showed similar results in an endotoxemia model and further identified that C5aR2 contributes to NLRP3 inflammasome activation and then HMGB1 release. In addition, our work utilized an MSU-induced peritonitis model to explore the role of C5aR2 in NLRP3 inflammasome activation in vivo. Our findings are consistent with earlier reports in which blocking the C5a receptor with an antagonist ameliorated inflammation (40,41). In summary, our work and others all suggest that C5aR2 contributes to harmful consequences in NLRP3-related disorders. Although one study found that C5a could suppress the NLRP3 inflammasome in murine macrophages in vitro and hypothesized that C5aR2 may contribute to the suppressive effect (18), we did not observe this phenomenon in C5aR2-deficient mice. Notably, C5aR1 plays a negligible role in NLRP3 inflammasome activation in this study (Fig. S2, A-C), which is in agreement with previous research (18). Moreover, the absence of both C5aR1 and C5aR2 had a similar effect on NLRP3 activation as C5aR2 deficiency alone. These results indicate that C5aR2 might have a dominant role in regulation of the NLRP3 inflammasome. There has been some research concerning the cross-talk of these two receptors. In CLP-induced sepsis, C5aR1 and C5aR2 have been shown to play a synergistic role in activating the NLRP3 inflammasome in cardiocytes (14). However, in CD4ϩ T cells, C5aR2 seems to inhibit intracellular C5aR1 (42). Interestingly, it has been reported that C5aR2 can heterodimerize with C5aR1, mediating either pro-or anti-inflammatory responses (43). This may be attributed to the balance of C5aR1 versus C5aR2 activation or rely on the concentrations of C5a. A high local concentration of C5a induces heterodimerization of C5aR1 and C5aR2, which facilitates anti-inflammatory cytokine production to protect from excessive inflammation (39,44). However, the underlying mechanisms of how this interplay happens and what it looks like in other immune cells remains to be explored in future studies.
PKR, a dsRNA-dependent protein kinase, revealed by our previous work to be an NLRP3 activator (22), is a surprising player to be involved in the C5a-C5aR2 axis. We observed that C5a-C5aR2 could promote the expression of PKR, almost at the transcription level. This amplifying effect depends on type I IFN signaling, which is critical for PKR expression, as PKR is an interferon-stimulated gene (45). In addition, we found that a MEK1/2 inhibitor as well as siRNA targeting MEK1 (Fig. 6A and Fig. S5, A and D) could reduce PKR expression promoted by C5a. Nevertheless, the details of MEK/ERK signaling and type I IFN signaling in C5a-mediated promotion of PKR expression need further investigation.
Abnormal activation of the NLRP3 inflammasome is related to inflammatory diseases and disorders, explaining why great attention has been paid to regulation of the NLRP3 inflammasome in these diseases (10, 46 -48). Our previous study found that PKR is involved in NLRP3 activation and identified ethyl pyruvate as an NLRP3 inflammasome inhibitor (22,49). In this study, we identified C5aR2 as an important modulator of NLRP3 inflammasome activation, identifying a potential target to treat NLRP3-associated inflammatory disorders (41). The benefits of targeting C5aR2 to treat inflammatory disorders

C5aR2 promotes NLRP3 activation
linked to NLRP3 relies on the tendency of C5aR2 to selectively regulate the activity of the NLRP3 inflammasome and have a minimal effect on other inflammatory responses (Fig. 2, A-C). However, it should be noted that C5aR2 expression and localization are possibly dynamic and cell-specific, undergoing continuous changes and being dependent on diverse milieus that make it difficult to be targeted with certain drugs or antibodies under pathological conditions (38,50). C5a, the natural ligand of C5aR2, could also bind to C5aR1 with high affinity and trigger C5aR1-dependent signal transduction, which probably has unknown effects on C5aR2-mediated inflammatory responses (16,26,51). Nevertheless, the latest discovery of C5aR2's role in inflammasomes has further enhanced our understanding of the mechanism of NLRP3 activation and might help to advance the treatment of NLRP3-related diseases.

Experimental procedures
Mice WT C57BL/6 mice (8 -10 weeks old) were purchased from Hunan SJA Laboratory Animal Co. Ltd. (Changsha, China). C5aR2 Ϫ/Ϫ mice in a C57BL/6 background were purchased from Model Animal Research Center (Nanjing, China). The PKR Ϫ/Ϫ mice were generous gifts from Dr. Kevin J. Tracey. The IFN␣␤R Ϫ/Ϫ mice were generous gifts from Dr. Jin Hou. All animals were kept under specific pathogen free conditions. Experimental groups were sex-matched and 8 -12 weeks of age. For gene knockout mice, WT littermate mice were used as controls. Animals were maintained in the Central South University Animal Facility. Studies were conducted in accordance with the Institutional Animal Care and Use Committee of Central South University.

Generation of elicited peritoneal macrophages
Mice were injected intraperitoneally with 3 ml of thioglycollate (3%). After 72 h, mice were sacrificed, and the elicited cells were harvested by peritoneal lavage with 15 ml of RPMI 1640 medium and resuspended in complete medium (RPMI 1640 medium supplemented with 10% FBS, 100 units/ml penicillin, and 100 g/ml streptomycin). For subsequent experiments, macrophages were seeded into 24-well plates (for ELISA or real time qPCR) or 6-well plates (for Western blotting) with a concentration of 1 ϫ 10 6 /ml, followed by replacement of medium 2 h later to remove nonadherent cells.

MSU-induced mouse peritonitis
C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mice (8 -10 weeks old and sexmatched) were injected intraperitoneally with 1 mg of MSU dissolved in 0.5 ml of sterile PBS. Mice were sacrificed 6 h later, and the peritoneal cavities were flushed with 5 ml of cold PBS. Peritoneal lavage fluids were collected, and cytokines were measured by ELISA after concentration using an Amicon Ultra 10 K filter (UFC900308) from Millipore. The pelleted cells were counted and analyzed for influx of neutrophils by cytometry. Trypan blue staining was used for exclusion of dead cells. The investigators were not blinded to allocation during experiments. The analysis was performed blindly by an independent researcher.

LPS-induced mouse endotoxemia
C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mice (8 -10 weeks old and sexmatched) were injected intraperitoneally with LPS (10 mg/kg body weight) for 12 h, and the serum levels of IL-1␤, TNF␣, and HMGB1 were measured by ELISA. Survival of C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mice subjected to 10 mg/kg LPS were monitored for 4 days. The investigators were not blinded to allocation during C5aR2 promotes NLRP3 activation experiments. The analysis was performed blindly by an independent researcher.

Transfection of the PKR expression vector to macrophages
We used a murine PKR expression vector described previously (22). For the rescue experiments, the recombinant PKR expression vector and the control vector were transiently transfected into peripheral macrophages of C5aR2 ϩ/ϩ and C5aR2 Ϫ/Ϫ mice using an Amaxa P3 Primary Cell 4D-Nucleofector X kit (Lonza) according to the manufacturer's instructions.

ELISA assay for cytokines
The levels of IL-1␤ and TNF␣ collected from plasma of mice or from the culture medium in vitro were determined using quantitative ELISA kits (eBioscience) according to the manufacturer's instructions. The released HMGB1 in the serum or supernatant was determined using a commercial ELISA kit (Shino-Test).

Statistical analysis
All statistical analyses were performed with GraphPad Prism software. Statistical significance was determined with twotailed unpaired Student's t test. Mouse survival data were plotted as Kaplan-Meier curves and compared by log-rank (Mantel-Cox) test. p Ͻ 0.05 was considered statistically significant, with increasing levels of confidence displayed as follows: *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001.